onnxruntime
webgpu: Extend FlashAttention decode path for any sequence length
#28389
Open
Open
webgpu: Extend FlashAttention decode path for any sequence length #28389
qjia7
wants to merge
1 commit into
main
from
webgpu-flash-attention-relax-subgroups
webgpu: Extend FlashAttention decode path to support any sequence length
435c4456
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub